- 



1 1 1 1 1 1 1 

 FEMALE BLUE MARLIN (N 



■86) > , ¥ « x " 



...» _ " 



- 



- 





«x * x x » x * 

 " "** "" »" x 



_ 



- 



X 



X *< x * 



- 



- 





X 



- 



- 



" 





- 



- 



1 1 I 1 I 1 1 





- 



. . " "^ . 



» . .. V . * 



-. i . < » ; 



X » * " x"" , " . 



"« x .»*«""« x X 



x * . "„ 



' 3.00 3.40 3.80 4.20 4.60 5.00 5.40 5.80 6.20 5.30 5.40 5.50 5.60 5.70 5.80 5.90 &00 6.10 

 0.24, , , , , . , , . , , , , . , . r 



0.08 

 en 



< 



9 o.oo 

 <n 



UJ 



IE 



-0.08 



FEMALE SLUE MARLIN (N=85) 



« «" x* 



■ . 1 1 



' i r - 



3.00 3.40 3.80 4.20 4.60 5.00 5.40 5.80 6.20 5.30 5.38 5.46 5.54 5.62 5.70 5.78 5.86 5.94 



n 1 1 1 r~ 



STRIPED MARLIN (N-53) 



1 ; i 1 \ \ ^ 1 1 ! r 



_i i i i i i i_ 



_i i i i i i i_ 



j, I , 1 1 1 1 , 1 1 1 1 1 , 1 1 L 



<!.00 2.40 2.80 3.20 3.60 4.00 4.40 4.80 5.20 4.90 5.00 5.10 5.20 5.30 5.40 5.50 5.60 5.70 

 LOG. WEIGHT (KG) LOG. FORK LENGTH (CM) 



Figure 2. — Plot of residuals from the log-linear model for female blue marlin with 86 and 85 samples and for 

 striped marlin with 53 samples. Weight was recorded in kilograms and fork length in centimeters. 



however, the elimination of this datum did not alter 

 the results significantly. When the trolling data were 

 divided into males and females, the error terms were 

 still not normally distributed. However, when the 

 above mentioned aberrant datum for the female data 

 was dropped from the calculations, the error terms 

 were normally distributed. The "F" test showed 

 that the relationship was highly significant, and the 

 relationship accounted for 939c of the variation in the 

 data. 



For large blue marlin (five relationships) and 

 striped marlin, the residuals about the regression line 

 were plotted against the dependent (weight) and in- 



dependent (fork length) variables in order to 

 evaluate the fit of the log-linear model. In every 

 case, the distribution of the residuals appeared as a 

 band along the axes; hence, the model appeared to fit 

 the data. The results for striped marlin and blue 

 marlin (trolling data for females with all data points 

 and with the one aberrant datum point dropped) 

 were representative of all the species plots. These 

 results are presented in Figure 2. The two plots for 

 the blue marlin indicated the effect of the aberrant 

 datum that was discussed earlier when the normality 

 of the residuals was tested. In spite of the residuals 

 not being normally distributed for all except two of 



132 



